Original Insight Summary

🤖 AI Data Moats & LLM Infrastructure

• Reddit ($RDDT) has a stronger AI-data moat than the market is pricing in. Early engineers from major LLMs reportedly viewed Reddit data as essential to model creation, and Reddit being the #1 most cited domain reinforces the idea that its corpus is core training and retrieval infrastructure, not just social content. 1

🇰🇷 Korea, NVIDIA & AI Supply Chain Positioning

• The old “Korea-passing” narrative looks like a useful lens for reading current AI supply-chain shifts. When Jensen Huang used to skip Korea while visiting Taiwan and China, Korean media framed it as Big Tech bypassing Korea. That history matters because it shows how executive travel patterns can become a proxy for where AI hardware leverage is really concentrating. 2

🧩 Advanced Packaging & Semiconductor Supply Chain

• Panel-level packaging is being framed as the next packaging endgame. The core takeaway: scale, substrate efficiency, and cost structure are becoming decisive as AI accelerators push packaging beyond traditional wafer-level constraints. 3

• TrendForce exposing the full CoPoS supply chain is a tell. Once the supplier map becomes visible, the market can start repricing upstream beneficiaries instead of treating advanced packaging as a black box. 4

❄️ Software Valuation & Earnings Quality

• Snowflake (#SNOW) delivered a strong FY2027 Q1 report, but the valuation debate is not clean because the company is still loss-making. The key swing factor is SBC; stock-based compensation will decide how much of the growth investors can actually underwrite in future multiples. 5